Characterization of enhanced monovalent and bivalent thrombin DNA aptamer binding using single molecule force spectroscopy

Abstract

Thrombin aptamer binding strength and stability is dependent on sterical parameters when used for atomic force microscopy sensing applications. Sterical improvements on the linker chemistry were developed for high-affinity binding. For this we applied single molecule force spectroscopy using two enhanced biotinylated thrombin aptamers, BFF and BFA immobilized on the atomic force microscopy tip via streptavidin. BFF is a dimer composed of two single-stranded aptamers (aptabody) connected to each other by a complementary sequence close to the biotinylated end. In contrast, BFA consists of a single DNA strand and a complementary strand in the supporting biotinylated part. By varying the pulling velocity in force-distance cycles the formed thrombin-aptamer complexes were ruptured at different force loadings allowing determination of the energy landscape. As a result, BFA aptamer showed a higher binding force at the investigated loading rates and a significantly lower dissociation rate constant, k(off), compared to BFF. Moreover, the potential of the aptabody BFF to form a bivalent complex could clearly be demonstrated.

Figure 1

Thrombin immobilization. (A) Topographical MAC mode image of a dense thrombin monolayer. Image size is 2.0 × 2.0 μm². A significantly increased indentation force was applied to remove thrombin molecules from the surface with a square of 500 × 500 nm². The overlaid cross-section profiles show a layer height of 2–3 nm. Scale bar length is 500 nm. (B) Schematic representation of thrombin aptamers (BFA and BFF) sensitive to fibrinogen binding site of thrombin (see Experimental section).

Figure 2

Single molecule force spectroscopy experiments. (A) The heterobifunctional cross-linker NHS-PEG₁₈-acetal (or NHS-PEG₁₈-aldehyde) was covalently bound to the aminopropyltrioxysilane-coated AFM tip through its NHS-ester function. Streptavidin was attached to the free aldehyde residue (after deprotection of the acetal group). Finally, the biotinylated thrombin-aptamer BFA or BFF was coupled to streptavidin. (B) Typical force-distance cycle for the specific interaction between a thrombin-aptamer on the tip and a thrombin molecule on the surface. The parabolic-like shape of the retrace curve (black line) corresponds to the dissociation of the bond between thrombin and the BFF aptamer. The specific unbinding events vanished after blocking the aptamer tip by adding free thrombin (insert). (C) Typical pdf profile of the most probable unbinding force observed for BFF in absence (solid line) and presence (dashed line) of free thrombin. The highest peak of the Gaussian fit for unblocked condition is found at ∼77 pN at a loading rate of 1440 pN/s. (D) Binding probability of BFA and BFF before and after tip block with thrombin shown for one representative tip for each aptamer. The number of unbinding events was quite comparable for BFA and BFF.

Figure 3

Dynamic force spectroscopy measurements. (A) Comparison of the dynamic force spectra of BFA and BFF. The most probable unbinding forces were plotted against the logarithm of the loading rate. The lowest forces were found for single binding interaction of BFF (solid green squares). The forces of stabilized thrombin-aptamer BFA (solid orange squares) were slightly higher compared to those found for double binding of BFF (open green triangles). Dynamic force spectroscopy data from BFA and single binding BFF were fitted against Eq. 2. The solid lines represent the acquired fit functions, respectively. Double binding BFF data (open green triangles) were in good agreement with the expectations of a Markovian binding model (Eq. 3). BFA revealed the lowest value for dissociation indicating a more proper fitting compared to homodimeric BFF. (B) Pdfs of the most probable unbinding forces expressed by a Gaussian fit supplemented with a force histogram for BFA and BFF at comparable loading rates, i.e., ∼8900 pN/s for BFA and 14,800 pN/s for BFF (pdfs correspond to the circled data points in (A)). The solid line plots in both panels indicate the pdfs built up by the sum of Gaussian fits over all forces. The dashed line plots represent the contributions of the single and double bindings to the distribution fitted by a simple Gaussian function. A prominent second force peak was only obtained in the case of BFF (dashed lines). (C) Validation of BFF double binding. The left panel shows a representative image of a surface with lower thrombin molecule density that was used to evaluate BFF binding characteristics; scale bar of 300 nm. By applying a higher indentation force the molecules were removed to the outer edges of the scratching area. BFF double binding, which was observed with a high probability on a dense thrombin layer (upper part in right panel), was significantly reduced when single thrombin molecules were probed (lower part in right panel). The Gaussian fitting procedure to the most probable unbinding forces gained on a dense thrombin layer indicated a probability of 44% for the occurrence of a single binding and 51% for the occurrence of a double binding (upper part of the right panel). In contrast thereto, the single binding occurrence was increased to 85% on single thrombin molecules, whereas the double binding occurrence was lowered to 15% (lower part of left panel).